Projekt 1 - faza 3
RafaĆ Cylwa, Dariusz Matlak, Grzegorz Ziajka
8 maja 2016
Short introduction
The aim of this report is to present the results regarding analysis of visitorsâ paths in Centrum Nauki Kopernik. Our analysis focus on three main issues: local path shapes, connections between path features and visiting multi stage exhibits and queue formation problem.
Global path shape
The following map shows âtypical pathsâ during visits. The darker the segment the more frequently visitors go from an exhibit to another exhibit.
It can be easily seen that the majority of visitors is very active in the early stage of their visits whereas in further sectors they donât use their cards and hardly no transition is visited signifficantly often.
Local path shapes
In order to find some local regularities in paths we used association rules mining techniques. We checked the following supposition: if a visitor visits endpoints of some pathlike part of the exhibition then he visits other exhibits in that segment. We focused on four groups. We present the results
1)stations cnk12, cnk16, cnk18, cnk20, cnk21, cnk24, cnk78a
2)cnk18, cnk24, cnk37, cnk38, cnk48
3)cnk22, cnk39, cnk49, cnk60, cnk72
4)cnk46a, cnk46b, cnk47, cnk75.
The graphs suggest that our hypothesis may be true. All found rules have rather high values of the lift coefficient. We believe that reorganization of the exhibition in more path-like pattern could make all stations looking equally attractive and avoid situations where a station is missed in many paths just because it is placed less encouraging place that other stations do.
Visualization
To visualize paths found above we draw them over a map of exhibits in CnK. Each of stations present on a path is market by a circle or a box filled with color consistent with path, so we have Lime for first rule, Pink for second, Red for third and Blue for last one. Exhibits that are close to the proposed path shape but that were omitted for some reason were marked with Orange.
